The iterative shrinkage thresholding algorithm reveals dynamic aging trajectories of human T lymphocytes via multidimensional spectral flow cytometry analysis

Immunosenescence is a fundamental hallmark of aging, characterized by differential susceptibility across immune cell lineages. T lymphocytes are particularly vulnerable; however, the distribution and dynamics of aging-associated immune markers across T cell subsets remain incompletely characterized. In this study, we enrolled 462 healthy individuals stratified into six groups spanning 20 to over 70 years. Using multiparametric spectral flow cytometry, we systematically characterized T lymphocyte subsets, capturing both immunophenotypic diversity and functional status. We identified a progressive age-associated decline in the frequencies of CD8⁺, γδ⁺, and Vδ2⁺ T cells, alongside an increase in CD4⁺ T cells. Aging was marked by the reduction of naive T cells and an expansion of terminally differentiated effector memory populations. Among all subsets, CD8⁺ T cells exhibited the greatest sensitivity to age-associated immunophenotypic remodeling, characterized by reduced expression of CD27 and CD28, and increased expression of senescence-associated surface markers including CD57 and KLRG1, together with elevated cytotoxic effector molecules including IFN-γ and granzyme B. Using these parameters, we defined 10 immune biomarkers through machine learning approaches that accurately predicted chronological age (R² = 0.81, P < 0.001). We further reconstructed the trajectory of T cell aging using ISTA-based sparse coding, which can capture aging-related immunophenotypic patterns in over 80% of individuals across age groups. Together, these findings demonstrate the dynamic remodeling of T lymphocytes across the human lifespan and provide a foundation for immune age modeling and risk stratification for unhealthy aging.